Go back to publications View external

LANTERN: A Closed-Loop Benchmark for VLM-Based Cooperative Driving with Temporally Grounded Warnings

September 2026 Yongshuo Liu, Xu Gao, Morui Zhu, Yongqi Zhu, Qi Chen, Deyuan Qu, Song Fu, Qing Yang arXiv preprint arXiv:2609.06368, 2026 (under review)
#Autonomous Driving #VLM #Benchmark

LANTERN is a closed-loop benchmark for evaluating how vision-language models respond to safety warnings during cooperative driving. It separates warning onset, hazard onset, warning termination, and post-hazard recovery across diverse scenarios, using paired test conditions to isolate the impact of warnings on driving performance.